Non-diabetic premenopausal women have a much lower incidence of cardiovascular disease (CVD) than men, but premenopausal women with diabetes appear to lose this advantage and may have a greater CVD risk than men. Interventions to reduce CVD risk are therefore of great potential benefit. Our hypothesis is that in premenopausal women with NIDDM, exercise training will reduce CVD risk by improving lipids and lipoproteins. Sixty women with NIDDM will participate in a six month program of exercise training (30 exercise and 30 nonexercising controls) with measurements of exercise performance and risk factors at three and six months to document the time course of changes. Degree of disease will be assessed by measuring insulin sensitivity and glycemic control Treadmill testing and physical activity questionnaire will evaluate exercise performance. Lipids and lipoproteins to be measured include total cholesterol, HDL-C and subfractions, LDL-C, VLDL-C and apolipoproteins A-1 and B. A subset of women will participate in focused studies to elucidate potential mechanisms for how exercise training may affect risk factors. For this purpose, lipid composition (LC) and cholesteryl ester transfer (CET) activity will be measured to determine whether exercise training reduces CET (accelerated in NIDDM) and improves LC. Reduction of cardiovascular risk factors in premenopausal women has important implications since atherosclerotic complications are the leading causes of death in persons with NIDDM.